Diagnosability of Linear Hybrid Systems with discrete dynamics modelled by Petri nets*

Abstract

This work is concerned with the diagnosability of Linear Hybrid Systems (LHS), defined as a family ℱ of linear systems (LS’s) and a safe State Machine Petri net (SMPN) that governs the switching among the LS’s. In this work, faults are defined as particular LS’s describing faulty dynamics, i.e. the set ℱ is split into normal LS’s and faulty LS’s. The approach herein presented combines the distinguishability property in LS’s and observability concepts from SMPNs to provide sufficient conditions for diagnosability.